Long term trends in unionization
- Unionization and coverage data
- Unionization rates declined among men, but remained steady among women
- Unionization rates declined in all provinces
- Unionization rates declined further in some industries in the 2000s, but rose in others
- Data sources, methods, and definitions
From 1981 to 2012, the overall unionization rate—defined as the proportion of all employees who are union members—declined. In this brief analysis, long-term trends in unionization rates are examined by gender, age group, province and industry using data from the Labour Force Survey and a collection of earlier data sources.
- From 1981 to 2012, Canada’s unionization rate declined from 38% to 30%; most of the decline took place during the 1980s and 1990s.
- The decline was observed among men, but not among women. The percentage of men in unionized jobs fell from 42% in 1981 to 29% in 2012; women’s unionization rate hovered around 30% during the same period.
- However, the relative stability in women’s unionization rates masks two offsetting trends. The rate moderately declined among women younger than 45, but rose among those aged 45 to 64. Among men, unionization rates declined in all age groups.
- Unionization rates declined in all provinces over the period. In 2012, Alberta had the lowest unionization rate, 22%; Newfoundland and Labrador, at 38%, and Quebec, at 37%, had the highest.
- The overall unionization rate has changed little in recent years, although some changes have occurred within industries. For example, the rate has declined in some goods-producing industries, such as manufacturing.
The decline in unionization is not a recent phenomenon. In Canada, the unionization rate fell steadily through the 1980s and 1990s, but remained relatively stable through the 2000s. However, trends have differed across population groups. In this analysis, unionization trends are examined over the 1981-to-2012 period to better understand changes by gender, province and age group.Note1 Industry trends over the period are also discussed.
This report also accompanies the release of a new series of CANSIM tables on union membership and coverage across several key characteristics, based on Labour Force Survey (LFS) data.
Since 1997, Statistics Canada’s LFS has collected information on union membership and coverage. Specifically, employed individuals are asked whether they are 1) a union member; 2) not a union member, but covered by a collective agreement; or 3) not a union member and not covered by a collective agreement.
Prior to 1997, the LFS did not collect information on union membership or coverage by a collective agreement. However, the 1984 Survey of Union Membership (SUM) and the Labour Market Activity Survey (LMAS), conducted from 1986 to 1990, collected information on both union membership and coverage. In addition, the 1981 Survey of Work History (SWH) collected information only on union membership.Note2 In this analysis, these sources are combined to create a consistent time series from 1981 onwards on unionization rates, based on all employed workers aged 17 to 64 (see Data sources, methods, and definitions). The unionization rate is defined as the proportion of employed workers who are union members.Note3
From 1981 to 2012, the overall unionization rate fell from 38% to 30%, a decline of 8 percentage points (Chart 1). The rate declined by 2 percentage points during the 1980s, by another 6 percentage points during the 1990s, and then remained fairly stable over the 2000s.
Not all groups saw a similar decline. In 1981, men had a much higher rate than women—42% versus 31%. However, over the next three decades the rate remained fairly stable among women but declined among men. As a result, by 2012 men had a slightly lower unionization rate than women (29% and 31%, respectively).
The decline in the unionization rate was more pronounced among younger workers (Table 1). In 1981, the difference in the unionization rate between employed people aged 17 to 24 and those aged 55 to 64 was 16 percentage points (26% and 42%); by 2012, that difference had widened to 21 percentage points (15% and 36%).
|17 to 24||26.4||19.8||18.4||12.5||13.8||14.8|
|25 to 34||39.8||39.7||34.7||25.6||27.6||28.4|
|35 to 44||42.0||46.1||42.9||34.6||31.2||31.2|
|45 to 54||41.7||44.4||44.6||41.5||38.3||35.9|
|55 to 64||41.9||44.7||41.6||36.4||36.4||36.0|
|17 to 24||29.2||23.2||19.9||13.5||14.6||15.5|
|25 to 34||43.3||42.2||37.1||25.0||25.9||26.0|
|35 to 44||46.1||50.7||45.6||36.0||30.6||29.2|
|45 to 54||47.8||50.5||49.9||43.0||38.5||35.1|
|55 to 64||48.6||49.5||48.0||38.5||35.9||34.7|
|17 to 24||23.1||16.3||16.8||11.3||13.0||14.2|
|25 to 34||34.7||36.6||32.0||26.4||29.5||31.0|
|35 to 44||36.3||40.3||39.9||33.2||31.8||33.4|
|45 to 54||32.9||36.2||38.2||40.0||38.1||36.7|
|55 to 64||29.9||36.9||31.7||33.8||36.8||37.3|
|Sources: Statistics Canada, Labour Force Survey, 1999, 2007 and 2012; Labour Market Activity Survey, 1989; Survey of Union Membership, 1984; Survey of Work History, 1981.|
However, the trends across age groups differed between men and women. Among men, the rate declined in all age groups, but was especially pronounced in the 25-to-34 and 35-to-44 groups. In both these age groups, the rate declined by about 17 percentage points over the 1981-to-2012 period. The decline among men aged 25 to 34 mostly occurred during the 1980s and 1990s; the decline among those aged 35 to 44 took longer to materialize, extending into the 2000s.
The unionization rate also declined among younger women, but to a lesser degree, largely because the declines of the 1980s and 1990s were offset by gains in the 2000s. In all, the unionization rate declined by 9 percentage points among those aged 17 to 24, and by 3 to 4 percentage points among those aged 25 to 44. Only older women saw rates increase over the 1981-to-2012 period: 4 percentage points among the 45-to-54 age group and 7 percentage points among those aged 55 to 64.
Unionization rates have long differed between provinces, and trends in the rates also varied across provinces from 1981 to 2012 (Chart 2). The most significant provincial findings were as follows:
- In 2012, Newfoundland and Labrador, 38%, and Quebec, 37%, had the highest unionization rates. These two provinces also had the highest rates in 1981: 45% in Newfoundland and Labrador; 44% in Quebec.
- As in 1981, the lowest rate in 2012 was in Alberta, 22%. Ontario had the second lowest rate in 2012 at 27%.
- From 1981 to 2012, unionization declined in all provinces, but the largest declines took place in British Columbia, -13 percentage points, and New Brunswick, -11 percentage points. While most provincial declines took place in the 1980s and 1990s, the decline in British Columbia continued into the 2000s.
- In contrast, Manitoba had the smallest decline over the period, -3 percentage points. Other provinces with more moderate declines included Saskatchewan and Nova Scotia, both -4 percentage points.
What factors explain the decline in unionization rates in Canada over the 1980s and 1990s? A previous studyNote4 contends that part of the decline in the unionization of younger men was owing to employment shifts from industries and occupations with higher unionization rates, such as construction and manufacturing, to industries and occupations with lower rates, such as retail and professional services. The study also found that unionization rose among older women in part because they became more concentrated in highly unionized industries such as health care, educational services and public administration.
Employment shifts did not explain everything, however. For instance, among men aged 25 to 34, it explained less than half of the total decline in unionization rates between 1981 and 1998. This means that changes within industries and occupations also played a role.Note5
Since 1999, the overall unionization rate has been stable at 30%, but some changes have occurred within industries (Table 2).Note6 Specifically:
- From 1999 to 2012, the unionization rate in goods-producing industries declined by 4 percentage points, but rose slightly, by less than 1 percentage point, in services-producing industries. Thus, in 2012, the unionization rate for services-producing industries, 31%, exceeded the rate for goods-producing industries, 27%; it was the reverse in 1999—30% in services and 31% in goods.
- Within goods-producing industries, three sectors saw even larger declines in unionization rates over the period: manufacturing, -7 percentage points; utilities -5 percentage points; and forestry, fishing, mining, oil and gas, -5 percentage points.
- Within services-producing industries, some industries increased their unionization rate—particularly management, administrative and other support, +5 percentage points; and public administration, +2 percentage points. However, these were offset by declines in others, such as information, culture and recreation, -3 percentage points; and transportation and warehousing, -2 percentage points.Note7
|1999||2007||2012||Change (1999 to 2012)|
|Forestry, fishing, mining, oil and gas||26.9||21.0||21.8||-5.1|
|Transportation and warehousing||42.7||41.1||40.5||-2.1|
|Finance, insurance, real estate and leasing||7.9||9.7||8.9||0.9|
|Professional, scientific and technical services||4.1||4.3||4.4||0.3|
|Management, administrative and other support||10.5||12.8||15.3||4.9|
|Health care and social assistance||53.0||53.6||53.6||0.7|
|Information, culture and recreation||27.6||25.8||25.0||-2.6|
|Accommodation and food services||6.6||7.8||6.7||0.1|
|Source: Statistics Canada, Labour Force Survey, 1999, 2007 and 2012.|
Unionization rates also varied across various occupations and other job characteristics. For instance, full-time work, longer job tenure, large firms, higher educational attainment, and better wages were all associated with higher unionization rates. Workers with higher educational attainment are also more likely to be unionized.
Details on unionization rates and coverage across these characteristics, as well as those discussed above, can be found in CANSIM tables 282-0220 to 282-0225. However, note that these tables provide information only from 1997 onwards, as they are based on the LFS.Note 8
Diane Galarneau is a senior analyst and Thao Sohn is an analyst in the Labour Statistics Division of Statistics Canada. The authors would like to thank Éric Fecteau of Labour Statistics Division for his assistance.
Morissette, René, Grant Schellenberg and Anick Johnson. 2005. "Diverging trends in unionization.� Perspective on Labour and Income. Vol. 6 no. 4. Statistics Canada Catalogue no. 75-001-X.
Riddell, Craig. 1993. "Unionization in Canada and the United States: A Tale of Two Countries.� In Small Differences That Matter: Labor Market and Income Maintenance in Canada and the United States. David Card and Richard B. Freeman (eds). University of Chicago Press.
Uppal, Sharanjit. 2011. "Unionization 2011.� Perspectives on Labour and Income. Vol. 23, no. 4. Statistics Canada Catalogue no. 75-001-X.
Data sources and methods
In this analysis, data from 1997 onwards is based on Labour Force Survey (LFS) data. The LFS surveys approximately 56,000 households monthly, and provides information on general labour market trends by industry, occupation, hours worked, participation rate and unemployment rate. For 1997 onwards, the LFS also provides data on the number of union members and the employees covered by a collective agreement. Union status is only collected for the employee's main job—the job for which the employee reports the most hours per week. The territories are excluded from the analysis because data for these regions were not always available.
Data for years prior to 1997 came from the 1981 Survey of Work History (SWH), the 1984 Survey of Union Membership (SUM) and Labour Market Activity Survey (LMAS), from 1986 to 1990. Both the SWH and SUM were based on the LFS sample design. Their target population was all persons residing in Canada, except inmates of institutions, full-time members of the armed forces, and residents of Yukon and the Northwest Territories, and persons living on Indian reserves.
The target population of the LMAS, a supplement to the LFS, was the non-institutionalized Canadian population aged 16 to 69, excluding persons living in Yukon, the Northwest Territories and persons living on Indian reserves.
To ensure comparability across the various data sources, this article focuses on paid workers aged 17 to 64. Unpaid family workers and self-employed workers are excluded.
Unionization rate: the number of employed individuals who are union members as a proportion of the total number of employed individuals.
Coverage rate (coverage by a collective agreement): the number of employed individuals (including both union members and non-unionized employees), who are covered by a collective agreement, as a proportion of the total number of employed individuals.
- See Uppal (2011) for previous analysis of unionization data.
- Prior to 1981, another time series called CALURA (Corporations and Labour Unions Returns Act) was also collected, dating back from the early 1960s. No longer collected since the mid-1990s, the CALURA series is not comparable to household survey information because it was based on firm-level data. Data from CALURA can be found in CANSIM tables 279-0024 to 279-0028.
- Another measure of union activity is the union 'coverage rate', defined as the proportion of employees covered by a collective agreement, whether they are union members or not. This information can be calculated only from 1984 onwards because the 1981 Survey of Work History only collected information about membership, not coverage. Since 1984, however, the difference between the unionization rate and the coverage rate has never exceeded 5 percentage points, indicating that the vast majority of employed people who are covered by collective agreements are, in fact, union members. In 2012, the unionization rate was 30%; the coverage rate was 32%.
- See Morissette, Schellenberg and Johnson (2005). "Diverging trends in unionization.�
- See Riddell (1993) for a complete discussion of other factors that may influence the unionization rate.
- Long-term comparisons across industries are not possible because of changes in industry classification systems. In this analysis, changes from 1999 to 2012 are examined across the North American Industrial Classification System categories. In Morissette, Schellenberg and Johnson (2005), changes between 1981 and 1998 were examined across the Standard Industrial Classification industries.
- The employment mix across industries also changed over the period. For example, the share of the employed population aged 17 to 64 in manufacturing industries decreased from 18% in 1999 to 12% in 2012. Both changes in the rates within industries and changes in the distribution of employment across industries can have an impact on overall rates.
- The CANSIM tables on unionization provide information from the Labour Force Survey, but are not related to earlier data sources.
- Date modified: